PROJECT SUMMARY/ABSTRACT Despite immense resources devoted to developing better treatments, glioblastoma multiforme (GBM) patients face a paucity of treatment options and universally have poor outcomes. With treatment, the median survival is fifteen months. Immunotherapy treatment likely holds the greatest potential to combat this terrible disease, and yet many questions remain regarding the relationship between the immune system and the brain, and in particular, the immune system and cancer of the brain. Which components of immune response are indispensable for the immune response to glioblastoma? Why does the immune system ultimately fail to control GBM? Further, how can we leverage the immune response to better treat glioblastoma? This project focuses on interrogating these aspects of glioblastoma, and to effectively do so begins with better mouse models, which is the central thrust of this project. The immune editing hypothesis postulates that cancer cells present peptides derived from their own mutated proteins, which can be recognized as foreign by the immune system. These mutations function as targets, which can mark the tumor for destruction by cytotoxic T cells. During the course of this process, the tumor evolves to become less immunogenic and persists in equilibrium with the immune system. Eventually, the tumor gives rise to less immunogenic targets and effectively becomes invisible to the immune system. At this point, the tumor escapes beyond the immune system?s control and presents clinically. How this process occurs in the central nervous system is less well understood. Human GBMs pose the additional challenge in that they carry low mutational burden, and hence have few mutations to target. Current mouse models of GBM either have high mutational burden, or low mutational burden but are for various reasons unsuitable to study tumor-immune system interactions. The model for this project will use lentivirus as a means to deliver both oncogenic proteins and Cre-recombinase to excise loxP flanked tumor-suppressors in a precise way, in immunologically mature mice, of a pure genetic background. This autochthonous, orthotopic pre-clinical model will more faithfully recapitulate the disease as it occurs in humans. This system will also be employed to develop a hypermutated model of recurrent GBM to establish the relationship between mutational burden and checkpoint blockade sensitivity in CNS tumors. These models will be superior for simulating the disease seen in humans because of their low mutational burden, because the genetic targeting is precise, because how the immune system will ?see? the tumor is similar to the human disease, and because of the autochthonous nature of tumor formation. These genetically engineered mouse models of GBM will answer fundamental questions regarding how the immune system interacts with GBM, will establish the relationship between mutational burden and checkpoint blockade sensitivity of CNS tumors, and will hopefully lead to development of more effective treatments.